On Tue, May 28, 2013 at 1:46 PM, Finn Thain <fthain@telegraphics.com.au> wrote:
On Sun, 26 May 2013, Finn Thain wrote:
On Sun, 26 May 2013, Laurent Vivier wrote:
Le 26/05/2013 02:30, Finn Thain a ecrit :
This works if I disable "EARLY_PRINTK". To do that I need to patch
Kconfig :
--- a/arch/m68k/Kconfig.debug
+++ b/arch/m68k/Kconfig.debug
@@ -11,7 +11,7 @@ config BOOTPARAM_STRING
depends on BOOTPARAM
config EARLY_PRINTK
- bool "Early printk" if EMBEDDED
+ bool "Early printk"
depends on MVME16x || MAC
default y
help
My bad. When I added "|| MAC" I didn't realise that EARLY_PRINTK == y by
default when EMBEDDED == n. You may want to send that patch in formal
style. Or maybe this?
- depends on MVME16x || MAC
+ depends on EMBEDDED && (MVME16x || MAC)
Or maybe both?
On second thoughts, Laurent's patch seems like the better one. I don't
think EARLY_PRINTK is particularly relevant to EMBEDDED. I have my doubts
about the "default y" though. What do you think, Geert?
Yeah, sounds reasonable. The "if EMBEDDED" made the question to appear
only if EMBEDDED is set, and the "default y" enabled it if no question was
asked, which contradicts the "You should normally say N here, unless you want
to debug such a crash." comment.
Your patch is indeed another option, but it just makes it more difficult for the
user to enable EARLY_PRINTK, and enabling EMBEDDED will overflow the
casual's user brain with silly questions about other options he does need.
I'll go with Laurent's version. Any formal submission with SoB?